INDEX

Explanations

mentions of countries, particularly Uganda, and potentially legal or political connections between them

oai_token-act-pair · gpt-3.5-turbo

references to specific African countries, particularly Uganda and Nigeria

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alty

-1.00

++++++++++++++++

-0.75

sburgh

-0.73

 bets

-0.71

ancies

-0.69

ellation

-0.69

mble

-0.68

earable

-0.68

Denver

-0.67

ablishment

-0.65

POSITIVE LOGITS

andan

0.96

 Nigeria

0.96

 Erit

0.95

 Niger

0.92

 Haram

0.92

 Cameroon

0.91

 Leone

0.90

 Sudan

0.89

 Congo

0.88

 Senegal

0.86

Activations Density 0.071%

mentions of countries, particularly Uganda, and potentially legal or political connections between them

references to specific African countries, particularly Uganda and Nigeria

No Comments

No Known Activations